AI Paper Atlas

Research map / Probabilistic and scientific modeling / Uncertainty and Gaussian processes

Variational Bayesian Decision-making for Continuous Utilities

Tomasz Kuśmierczyk, Joseph Sakaya, Arto Klami

NeurIPS 2019 · Probabilistic and scientific modeling › Uncertainty and Gaussian processes · approach: Simulation-based inference · 6 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
calibrate with LCVI instead of standard mean-field VIEight schools model · Eight schools, tilted loss q=0.2empirical risk reduction: 1\%Improved
calibrate with LCVI instead of standard VIprobabilistic matrix factorization · Last.fmempirical risk reduction: 1-4\%Improved
use EM with numerical optimization of decisions instead of joint optimization (ablation)LCVI · Last.fm, squared loss, M90speed and accuracyHurt
EM with Bayes estimator vs joint optimization (ablation)LCVI · Last.fm, squared loss, M90accuracy and costNo clear effect
use LCVI instead of standard VIVI · Last.fmcomputational cost: 10 times fasterHurt
increase M (loss-to-utility cutoff quantile) to large values (ablation)LCVI · Last.fmrisk reductionHurt
set M between 50th and 95th loss quantile (ablation)LCVI · Last.fmrisk reduction: 70\%Improved
exponential instead of linearized loss-to-utility transformation (ablation)LCVI · Last.fmrisk reductionNo clear effect
calibrate directly for utility instead of linearized loss (ablation)LCVI · u=exp(-(h-y)^2)risk reductionImproved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy